UC Berkeley Frank Coyle：Agent 护栏应插入循环步骤之间
这段演讲把 Agent 安全从'最后兜底'改成'步步设卡'，讲得很实操：工具输入验 schema、输出做检查、结果看合不合理。正在写 Agent 工作流的同学可以直接照着加检查点。
UC Berkeley 的 Frank Coyle 在 AI Engineer Code Summit Day 1 演讲中提出，护栏不应只在 Agent 循环结束后统一校验，而应嵌入循环每一步之间。
具体做法包括三类检查点：验证工具输入是否符合 schema 与预期范围、在结果回传前检查模型输出、以及判断最终结果是否在语义上说得通。这样每次工具调用前后都形成独立的防线，出错时能在循环内部被拦截，而不是等到整轮任务跑完才发现问题。
对开发者而言，这种分步校验的方式可以直接套用到现有 Agent 框架里，减少幻觉调用和参数错误在多步任务中的累积。
Guardrails belong between the steps of the agent loop. @coyle_frankp of @UCBerkeley walks through where to validate tool inputs, inspect outputs, and check whether the result makes sense. From Day 1 at AI Engineer Code Summit.
This report is categorized under AITOP's "How-to" desk, focusing on US frontier AI labs and Silicon Valley technology leaders. Within our cross-regional analytical framework, it directly maps to AI developer toolchains, coding agents, and engineering workflows. Key entities and technical topics under continuous tracking include Agent, 护栏, AI Engineer Code Summit, 提示词工程.
Source attribution: curated from AI Engineer (@aiDotEngineer).
